[ NOTE 📌] about working with the PonyV6 model.

Weight: 0:8-1

• Weight: 0:8 - in my opinion, this allows you to follow the text more accurately in promt, but the style is not very accurately conveyed.

• Weight: 0:95 - almost perfectly conveys the style.

• Weight: 0:98 - a result that satisfies me in terms of style transfer.

⚠️ Dear friends, do not be offended if your generated image that does not match the style is removed from the examples.

ANDXL v1.0 - artist's style:

Trigger: andxl

Weight: 0:8

1. Prompt: score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up,

2. Negative prompt:

   1. none

   2. 3D, simple background, score_6, score_5, score_4, censored, monochrome, text, watermark,(worst quality, low quality), blurry face, clipping, bad face, deformed hands, (bad hands), bad feet, extra toes, extra fingers, heels, strabismus.

TEMXL v1.0 - weight: 0:8

• Prompt: score_9, score_8_up, score_7_up, score_6_up,

• Negative prompt: none - any negative word can ruin the style.

  Note: for some reason, the image generation on the site in this style does not match what you see.

SPMXL v1.0

• In certain situations, employ a single quality tag of score_9.

  Negative prompt:

• score_5_up, score_4_up, 3d

• score_5_up, score_4_up, 3d, blender, (outline:2)

  In some cases, avoid using negative words to convey style.

INCNXL v1.0 - this is one of the branches of Incase

• This is one of the branches of Incase, which was created with different dim and alpha parameters to test how the size of the output data affects after training.

• There are 3 files of different sizes in the archive.

  Negative prompt:

• score_6_up, score_5_up, score_4_up, 3d

• 3d, blender